Zhang et al. BMC Cancer 2014, 14:664 http://www.biomedcentral.com/1471-2407/14/664
RESEARCH ARTICLE
Open Access
Prognostic and therapeutic significance of ribonucleotide reductase small subunit M2 in estrogen-negative breast cancers Hang Zhang1†, Xiyong Liu2†, Charles D Warden3, Yasheng Huang2, Sofia Loera4, Lijun Xue4, Suzhan Zhang1, Peiguo Chu4, Shu Zheng1* and Yun Yen2*
Abstract Background: Ribonucleotide reductase (RR) is an essential enzyme involved in DNA synthesis. We hypothesized that RR subunit M2 (RRM2) might be a novel prognostic and predictive biomarker for estrogen receptor (ER)-negative breast cancers. Methods: Individual and pooled survival analyses were conducted on six independent large-scale breast cancer microarray data sets; and findings were validated on a human breast tissue set (ZJU set). Results: Gene set enrichment analysis revealed that RRM2-high breast cancers were significantly enriched for expression of gene sets that increased in proliferation, invasiveness, undifferentiation, embryonic stem/progenitor-like phenotypes, and poor patient survival (p < 0.01). Independent and pooled analyses verified that increased RRM2 mRNA levels were associated with poor patient outcome in a dose-dependent manner. The prognostic power of RRM2 mRNA was comparable to multiple gene signatures, and it was superior to TNM stage. In ER-negative breast cancers, RRM2 showed more prognostic power than that in ER-positive breast cancers. Further analysis indicated that RRM2 was a more accurate prognostic biomarker for ER-negative breast cancers than the pathoclinical indicators and uPA. A new RR inhibitor, COH29, could significantly enhance the chemosensitivity to doxorubicin in ER-negative MDA-MB-231 cells, but not in ER-positive MCF-7 cells. Conclusion: RRM2 appears to be a promising prognostic biomarker and therapeutic target for ER-negative breast cancer patients. Keywords: Ribonucleotide reductase, Breast cancer, ER-negative, Prognostic biomarker
Background Over a million new cases of breast cancer are diagnosed and ~400,000 deaths occur annually [1]. Breast cancer is a heterogeneous disease that has variable gene expression and different outcomes that cannot be predicted by pathologic grade or clinical stages [2,3]. A comprehensive gene expression signature has identified four major molecular subtypes: luminal A, luminal B, HER2-enriched and basal* Correspondence: [email protected]; [email protected]. † Equal contributors 1 Cancer Institute, Second Affiliated Hospital, Zhejiang University School of Medicine, 310009 Hangzhou, Zhejiang, China 2 Department of Molecular Pharmacology, City of Hope National Medical Center and Beckman Research Institute, 1500 E. Duarte Road, 91010 Duarte, CA, USA Full list of author information is available at the end of the article
like breast cancers, which are largely comprised of the triple negative breast cancer (TNBC) subtype. Each subtype has a distinct clinical behavior and response to therapy [4]. Among the four, the TN and HER2-enriched subtypes are considered to be ER-negative, and account for 15-17% and 15-20% of invasive breast carcinomas respectively [5]. Recently, therapies that specifically target the HER2 receptor have significantly increased the survival of patients with HER2-enriched breast cancers, and PARP (poly-ADP ribose polymerase) inhibition holds some promise as a targeted therapy for TNBC [6,7]. Gene signatures that include the 70-gene signature [8], 21-gene recurrence score (commercially developed as Oncotype Dx) [9], PI3K signature [10], core serum response signature (CSR) [11], and the grade signature [12], have been developed to
© 2014 Zhang et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Zhang et al. BMC Cancer 2014, 14:664 http://www.biomedcentral.com/1471-2407/14/664
predict the survival of breast cancer patients, and have been used to predict the outcome for ER-positive patients. A 7-gene immune response module (IRM) [13] and a HER2-derived prognostic predictor (HDPP, 158 genes) [14] were developed to identify ER-negative breast cancers that were associated with poor prognosis. Multiple-gene-based signatures potentially enhance the accuracy of prediction. However, their disadvantages include higher cost and lack of specific targets for chemotherapeutic agent selection. Ribonucleotide reductase (RR) is a rate-limiting enzyme essential for DNA replication. RR catalyzes the formation of 2′-deoxyribonucleoside diphosphates from corresponding ribonucleotide diphosphates [15]. Human RR is a heterodimeric tetramer that consists of two large RRM1 subunits and two small subunits, RRM2 and/or RRM2B [16]. RRM2 overexpression induces cell proliferation and invasion [17,18] and is positively correlated with higher grade and muscularis propria invasion in bladder [19] and gastric cancers [20]. Previously, we showed that high RRM2 expression is a prognostic indicator for poor survival for several kinds of cancers, including lung and colorectal cancers [21-23]. RRM2 enhances the invasive capacity of pancreatic adenocarcinoma cells in a NF-kappaB-dependent manner [24], and RRM2 overexpression can increase angiogenesis by down-regulating thrombspondin-1 and increasing production of VEGF [25]. Knocking down RRM2 also sensitizes cancer cells to the cytotoxic effects of the nucleoside analogs [26]. The RRM2 gene (mRNA) has been included in some gene signatures, such as the 3D-culture signature [27] and Five-gene Molecular Grade Index (MGI) [28,29], and the prognostic significance of RRM2 mRNA has been evaluated [30,31]. Independent prognostic significance of RRM2 protein needs to be further investigated, and the prognostic performance of RRM2 (mRNA and protein) also needs to be compared with other existing breast cancer biomarkers or gene signatures. Especially, the prognostic value of RRM2 in ER-negative breast cancers is worth further evaluation. In this study, we evaluated the prognostic value of RRM2 by analyzing seven independent breast cancer data sets, both individually and in a pooled manner. RRM2 expression was significantly correlated with poor survival in a dose-dependent manner, particularly for ER-negative breast cancers. Moreover, the prognostic value of RRM2 was comparable to the 70-gene signature and 21-gene recurrence score for breast cancer overall, and superior to these biomarkers and pathoclinical indicators for ER-negative breast cancers.
Methods Patients Microarray data sets
A total of 10 published microarray data sets including: Ivshina (GSE4922), Chin (E-TABM-158), Wang (GSE2034),
Page 2 of 16
Pawitan (GSE1456), Desmedt (GSE7390), Expo (GSE2109), Huang [32], Bild (GSE3143), Sortiriou (GSE2990) and NKI [33] with clinical annotations were downloaded from the combined microarray dataset BRAVO (Biomarker recognition and validation on-line). The Expo set was excluded because it lacks survival data. Based on BLAST results, the data sets of Huang, Bild and Sortiriou were excluded because of using low similarity probes (36922_at) (Additional file 1: Figure S1A). The NKI set was selected for RRM2 prognostic evaluation because the probe (Agilent Technologies) for RRM2 also is 100% similarity to sequence of > gi|260064012| ref|NM_001165931.1|, and it contains information of most gene signatures’ classification. Finally, six independent microarray data sets include Ivshina [12], Chin [34], Wang [35], Pawitan [36], Desmedt [37] and NKI [33] were chosen for this study. The summarized microarray data sets are shown in Additional file 2: Table S1. The significant correlation between signals yielded from two RRM2 probes (201890_at and 209773_s_at) in the Ivshina set was shown in Additional file 1: Figure S1B (R2 = 0.736, p < 0.001). Neither 201890_at nor 209773_s_at correlated with the expression of RRM2B or RRM1 (Additional file 1: Figure S1C to 1E), indicating that the signal was specific for RRM2 expression. ZJU set
The protocol for the use of human tissues was reviewed and approved by the Institutional Review Board (IRB) of the Second Affiliated Hospital of Zhejiang University (ZJU) (Zhejiang, China). Prior to the study, all patients gave their written informed consent to allow us to use leftover tissue samples for scientific research. All eligible participants had received modified radical mastectomy and the primary tumor samples were obtained from surgical specimens. The exclusion criteria were: 1) no informed consent obtained, 2) multiple cancers, 3) lack of histological diagnosis, and 4) no follow-up information. After applying the selection criteria, a total of 175 breast cancer patients who were diagnosed from 2002 to 2006 were enrolled in the ZJU set (Additional file 2: Table S2). All patients recruited were Chinese females with a median age of 50.0 years (range 20–84 years). The pathoclinical and demographic data was collected by reviewing the hospital records. Of the 175 patients, 46 patients with local advanced breast cancer( Stage IIIA, IIIB and IIIC) received neo-adjuvant chemotherapy before surgery, 140 patients with positive lympho node or aggressive potential(Including: Age
RESEARCH ARTICLE
Open Access
Prognostic and therapeutic significance of ribonucleotide reductase small subunit M2 in estrogen-negative breast cancers Hang Zhang1†, Xiyong Liu2†, Charles D Warden3, Yasheng Huang2, Sofia Loera4, Lijun Xue4, Suzhan Zhang1, Peiguo Chu4, Shu Zheng1* and Yun Yen2*
Abstract Background: Ribonucleotide reductase (RR) is an essential enzyme involved in DNA synthesis. We hypothesized that RR subunit M2 (RRM2) might be a novel prognostic and predictive biomarker for estrogen receptor (ER)-negative breast cancers. Methods: Individual and pooled survival analyses were conducted on six independent large-scale breast cancer microarray data sets; and findings were validated on a human breast tissue set (ZJU set). Results: Gene set enrichment analysis revealed that RRM2-high breast cancers were significantly enriched for expression of gene sets that increased in proliferation, invasiveness, undifferentiation, embryonic stem/progenitor-like phenotypes, and poor patient survival (p < 0.01). Independent and pooled analyses verified that increased RRM2 mRNA levels were associated with poor patient outcome in a dose-dependent manner. The prognostic power of RRM2 mRNA was comparable to multiple gene signatures, and it was superior to TNM stage. In ER-negative breast cancers, RRM2 showed more prognostic power than that in ER-positive breast cancers. Further analysis indicated that RRM2 was a more accurate prognostic biomarker for ER-negative breast cancers than the pathoclinical indicators and uPA. A new RR inhibitor, COH29, could significantly enhance the chemosensitivity to doxorubicin in ER-negative MDA-MB-231 cells, but not in ER-positive MCF-7 cells. Conclusion: RRM2 appears to be a promising prognostic biomarker and therapeutic target for ER-negative breast cancer patients. Keywords: Ribonucleotide reductase, Breast cancer, ER-negative, Prognostic biomarker
Background Over a million new cases of breast cancer are diagnosed and ~400,000 deaths occur annually [1]. Breast cancer is a heterogeneous disease that has variable gene expression and different outcomes that cannot be predicted by pathologic grade or clinical stages [2,3]. A comprehensive gene expression signature has identified four major molecular subtypes: luminal A, luminal B, HER2-enriched and basal* Correspondence: [email protected]; [email protected]. † Equal contributors 1 Cancer Institute, Second Affiliated Hospital, Zhejiang University School of Medicine, 310009 Hangzhou, Zhejiang, China 2 Department of Molecular Pharmacology, City of Hope National Medical Center and Beckman Research Institute, 1500 E. Duarte Road, 91010 Duarte, CA, USA Full list of author information is available at the end of the article
like breast cancers, which are largely comprised of the triple negative breast cancer (TNBC) subtype. Each subtype has a distinct clinical behavior and response to therapy [4]. Among the four, the TN and HER2-enriched subtypes are considered to be ER-negative, and account for 15-17% and 15-20% of invasive breast carcinomas respectively [5]. Recently, therapies that specifically target the HER2 receptor have significantly increased the survival of patients with HER2-enriched breast cancers, and PARP (poly-ADP ribose polymerase) inhibition holds some promise as a targeted therapy for TNBC [6,7]. Gene signatures that include the 70-gene signature [8], 21-gene recurrence score (commercially developed as Oncotype Dx) [9], PI3K signature [10], core serum response signature (CSR) [11], and the grade signature [12], have been developed to
© 2014 Zhang et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Zhang et al. BMC Cancer 2014, 14:664 http://www.biomedcentral.com/1471-2407/14/664
predict the survival of breast cancer patients, and have been used to predict the outcome for ER-positive patients. A 7-gene immune response module (IRM) [13] and a HER2-derived prognostic predictor (HDPP, 158 genes) [14] were developed to identify ER-negative breast cancers that were associated with poor prognosis. Multiple-gene-based signatures potentially enhance the accuracy of prediction. However, their disadvantages include higher cost and lack of specific targets for chemotherapeutic agent selection. Ribonucleotide reductase (RR) is a rate-limiting enzyme essential for DNA replication. RR catalyzes the formation of 2′-deoxyribonucleoside diphosphates from corresponding ribonucleotide diphosphates [15]. Human RR is a heterodimeric tetramer that consists of two large RRM1 subunits and two small subunits, RRM2 and/or RRM2B [16]. RRM2 overexpression induces cell proliferation and invasion [17,18] and is positively correlated with higher grade and muscularis propria invasion in bladder [19] and gastric cancers [20]. Previously, we showed that high RRM2 expression is a prognostic indicator for poor survival for several kinds of cancers, including lung and colorectal cancers [21-23]. RRM2 enhances the invasive capacity of pancreatic adenocarcinoma cells in a NF-kappaB-dependent manner [24], and RRM2 overexpression can increase angiogenesis by down-regulating thrombspondin-1 and increasing production of VEGF [25]. Knocking down RRM2 also sensitizes cancer cells to the cytotoxic effects of the nucleoside analogs [26]. The RRM2 gene (mRNA) has been included in some gene signatures, such as the 3D-culture signature [27] and Five-gene Molecular Grade Index (MGI) [28,29], and the prognostic significance of RRM2 mRNA has been evaluated [30,31]. Independent prognostic significance of RRM2 protein needs to be further investigated, and the prognostic performance of RRM2 (mRNA and protein) also needs to be compared with other existing breast cancer biomarkers or gene signatures. Especially, the prognostic value of RRM2 in ER-negative breast cancers is worth further evaluation. In this study, we evaluated the prognostic value of RRM2 by analyzing seven independent breast cancer data sets, both individually and in a pooled manner. RRM2 expression was significantly correlated with poor survival in a dose-dependent manner, particularly for ER-negative breast cancers. Moreover, the prognostic value of RRM2 was comparable to the 70-gene signature and 21-gene recurrence score for breast cancer overall, and superior to these biomarkers and pathoclinical indicators for ER-negative breast cancers.
Methods Patients Microarray data sets
A total of 10 published microarray data sets including: Ivshina (GSE4922), Chin (E-TABM-158), Wang (GSE2034),
Page 2 of 16
Pawitan (GSE1456), Desmedt (GSE7390), Expo (GSE2109), Huang [32], Bild (GSE3143), Sortiriou (GSE2990) and NKI [33] with clinical annotations were downloaded from the combined microarray dataset BRAVO (Biomarker recognition and validation on-line). The Expo set was excluded because it lacks survival data. Based on BLAST results, the data sets of Huang, Bild and Sortiriou were excluded because of using low similarity probes (36922_at) (Additional file 1: Figure S1A). The NKI set was selected for RRM2 prognostic evaluation because the probe (Agilent Technologies) for RRM2 also is 100% similarity to sequence of > gi|260064012| ref|NM_001165931.1|, and it contains information of most gene signatures’ classification. Finally, six independent microarray data sets include Ivshina [12], Chin [34], Wang [35], Pawitan [36], Desmedt [37] and NKI [33] were chosen for this study. The summarized microarray data sets are shown in Additional file 2: Table S1. The significant correlation between signals yielded from two RRM2 probes (201890_at and 209773_s_at) in the Ivshina set was shown in Additional file 1: Figure S1B (R2 = 0.736, p < 0.001). Neither 201890_at nor 209773_s_at correlated with the expression of RRM2B or RRM1 (Additional file 1: Figure S1C to 1E), indicating that the signal was specific for RRM2 expression. ZJU set
The protocol for the use of human tissues was reviewed and approved by the Institutional Review Board (IRB) of the Second Affiliated Hospital of Zhejiang University (ZJU) (Zhejiang, China). Prior to the study, all patients gave their written informed consent to allow us to use leftover tissue samples for scientific research. All eligible participants had received modified radical mastectomy and the primary tumor samples were obtained from surgical specimens. The exclusion criteria were: 1) no informed consent obtained, 2) multiple cancers, 3) lack of histological diagnosis, and 4) no follow-up information. After applying the selection criteria, a total of 175 breast cancer patients who were diagnosed from 2002 to 2006 were enrolled in the ZJU set (Additional file 2: Table S2). All patients recruited were Chinese females with a median age of 50.0 years (range 20–84 years). The pathoclinical and demographic data was collected by reviewing the hospital records. Of the 175 patients, 46 patients with local advanced breast cancer( Stage IIIA, IIIB and IIIC) received neo-adjuvant chemotherapy before surgery, 140 patients with positive lympho node or aggressive potential(Including: Age
